Generative AI vs. conversational AI and the impact on customer experience

Gartner recently released poll results showing that 38% of respondents consider customer experience/retention as their primary focus of generative AI investments. That was number one, ahead of revenue growth (26%), cost optimization (17%), and business continuity (7%). That’s a big deal – especially considering that in 2022, the CMSWire State of Digital Customer Experience report found that a quarter of respondents said they had no AI applications in their CX toolset. None.

Clearly generative AI is taking center stage right now, but there’s a different form of AI that’s been on the scene for a while – conversational AI.

In this blog article, we’ll talk about the AI that’s been at the center of automating contact centers and improving customer experience strategies (conversational AI) as well as the new kid on the block grabbing all the headlines (generative AI) and what each means for your CX plans.

Conversational AI and Generative AI Defined

First off, let’s define what we mean when we talk about conversational and generative AI and break down these terms:

Generative AI involves programming a computer to replicate a human mind in order to create new content. The dominant style of generative AI is based on the neural network, which is an estimation of how we think brain works. Generative AI takes data from a training set and then generates new data based on the patterns and characteristics of the training set.

Conversational AI, on the other hand, is designed to engage in back-and-forth interactions, like a conversation, with humans or other machines in a natural language. Conversational AI can be used to collect information, accelerate responses, and augment an agent’s capabilities.

3 Ways Conversational AI Is Transforming the Customer Experience

Conversational AI has actually been the backbone for many of the advances in customer experience that we take for granted today. Here are some of the main places conversational AI shows up in customer experience:

Chatbots

There are simple chatbots and there are advanced chatbots; the latter is powered by conversational AI. Traditional chatbots are rules-based and use a set script to respond to customer inquiries. If a customer asks a question in an unexpected way, the bot is easily stumped. Conversational chatbots, on the other hand, have an expanded ability to engage beyond their programming. Instead, they use a type of machine learning called Natural Language Processing (NLP) to recognize speech and imitate human interactions. Conversational chatbots can handle complex inquiries, operate across multiple channels, and actually learn through interactions over time.

Interactive Voice Response (IVR) Systems

We’re all familiar with calling a toll-free number and then being asked to select from a limited set of choices. That’s an old-school IVR system and it has a lot of the same problems as traditional chatbots – specifically that it can’t recognize an input outside of its scripted responses. Now IVR systems are getting a huge boost from conversational AI. With natural language processing (NLP), IVR systems can recognize conversational language and provide more accurate and personal responses. This technology also means that an IVR doesn’t need to include a long and complicated menu. Instead, customers can just say why they’re calling and be given the appropriate response or be routed to the right agent.

Sentiment Analysis

How does a chatbot know when a person is upset or dissatisfied with their service? Sentiment analysis. This conversational AI tool can actually pick up on verbal or textual queues to predict the customer’s emotional demeanor and either use tools to de-escalate the situation or transfer the customer to a live agent. Here’s an example: A customer’s Wi-Fi connection is down so he contacts his internet provider via chatbot to see what the problem is. It turns out that there is an outage in the area and crews are working to fix it. With sentiment analysis, the bot can understand that the customer isn’t satisfied with this answer, so the bot proactively offers the customer a discount on his bill. Suddenly the customer isn’t so dissatisfied anymore.

3 Opportunities for Generative AI to Impact Customer Experience

We really don’t see the robots taking over our contact center agents’ jobs – rather we think they’ll make the jobs better by giving them the tools to work with more competence and confidence. Our data shows that top-performing contact center agents have lower attrition rates than bottom-performing agents. So, one of the ways to stop the revolving door is to turn every agent into a top performer. Here’s how that plays out with generative AI:

More Robust Knowledge Centers

We joke that if you walk into a contact center and see agents’ desks plastered with sticky notes, you have a knowledge center problem. Agents are getting asked some of the same questions all the time so they jot the answers on sticky note, so they’ll be ready when the question inevitably arises. Knowledge centers powered by machine learning already do a lot to alleviate this problem by delivering answers to agents via tools in their contact center technology. However, with generative AI, this can get even better. Using existing knowledge bases, manuals, FAQs, case notes or other guides, generative AI can consume all of that content and use it to generate answers to just about any question an agent might receive.

Suggested Responses

Canned responses can be beneficial, but they’re limited in the number that can be effectively used by associates, creating a ceiling to productivity. And scripting voice calls is only effective in very specific, repeatable scenarios. Suggested response tools will use transcribed data from historical contacts to learn and predict the best response to a customer question or statement. With generative AI, this type of tool will be available in both chat and voice channels.

Real-time Coaching

Traditional coaching processes that leverage post-call quality or analytics processes require a long duration between the interaction and a coachable event, reducing the effectiveness that immediate feedback can provide. With generative AI and natural language processing, contact center agent supervisors will be able to provide coaching and quality suggestions in real time. In-the-moment feedback can provide benefits in new associate ramp-up time, quality, compliance, and conversion rates.

Where Do We Go from Here?

These applications are just the tip of the iceberg when it comes to both conversational and generative AI and we see many opportunities for advancements in both technologies. Technological innovations are exciting, but they’re only as good as the people and systems that support them. So before going all in on any kind of technology, we’d encourage you to do your homework and if you’re not an AI or CX expert, work with someone who is. Just because you can easily incorporate AI into your CX strategy, doesn’t mean you’ll get the results you want without strong design and expertise to back it up.